ord_nms: NMS with randomization test and stepdown dimensionality...

In phytomosaic/ecole: ecole: School of Ecology Package

Description

Nonmetric multidimensional scaling with randomization test for stopping rules and stepdown dimensionality selection.

Usage

ord_nms(
  x,
  autopilot = c("quick", "medium", "slow"),
  method = "bray",
  parallel = parallel::detectCores() - 1,
  ...
)

Arguments

x	array, where rows = SUs and cols = variables such as environmental or traits values.
autopilot	character, one of three possible thoroughness options.
method	dissimilarity index, passed to 'vegan::vegdist'.
parallel	number of cores to use in parallel processing. Set 'NULL' to avoid parallel processing.
...	further arguments passed to 'vegan::metaMDS'.

Details

NMS with stopping rules for dimensionality based on PC-ORD (McCune and Mefford 2017). This is time-consuming, use parallel processing where possible.

Value

List containing items returned by 'vegan::metaMDS' for the selected model, appended with further items:
- pval: permutation p-values for each dimensionality.
- stress_real_vs_rnd: stress values, where each column = dimensionality, and each row = either the real-data stress value (first row only) or the randomized-data stress value (successive rows).

References

McCune, B., and M. J. Mefford. 2017. PC-ORD. Multivariate Analysis of Ecological Data. Version 7. MjM Software Design, Gleneden Beach, OR.

See Also

metaMDS

Examples

## Not run: 
data(smoky)
m1 <- ord_nms(smoky$spe, 'quick')
m2 <- ord_nms(smoky$spe, 'medium')
m3 <- ord_nms(smoky$spe, 'slow')

## End(Not run)

phytomosaic/ecole documentation built on Jan. 2, 2022, 11:24 p.m.